7736
J. Org. Chem. 1997, 62, 7736-7743
Enzymatic Desymmetrization of Prochiral 2-Substituted-1,3-propanediols: A Practical Chemoenzymatic Synthesis of a Key Precursor of SCH51048, a Broad-Spectrum Orally Active Antifungal Agent Brian Morgan,*,† David R. Dodds,† Aleksey Zaks,† David R. Andrews,‡ and Ricardo Klesse‡ Schering-Plough Research Institute, Biotransformations Group, K-15-1/1800, 2015 Galloping Hill Road, Kenilworth, New Jersey 07033-0539, and Chemical Process R & D, 1011 Morris Avenue, Union, New Jersey 07083 Received June 3, 1997X
Two examples of a practical enzymatic desymmetrization of a 2-substituted-1,3-propanediol and their application to the synthesis of SCH51048, a broad-spectrum orally active antifungal, are described. In each case, enzymatic catalysis under both hydrolytic and transesterification conditions is described. In the first example the key intermediate, the R,S-monoester of triol 6, was obtained via Amano Lipase AK catalyzed hydrolysis of the dibutyrate 11b, or Novo SP435 catalyzed acetylation of triol 6. In the second example, desymmetrization of diol 13a using Novo SP435 or of dibutyrate 13c using Amano Lipase CE furnished the S-monoester (S)-14b,c, a key intermediate in a new efficient synthesis of SCH51048. Optimization of the Novo SP435 acetylation of diol 13a and the scaleup of the reaction is also described. Introduction SCH51048 (1) was identified as a potential antifungal agent with efficacy in the treatment of systemic Candida and Aspergillus infections in both normal and immunocompromised models.1 In common with other azole antifungals (Chart 1), SCH51048 (1) contains the heterocyclic ring, a dihalogenobenzene ring and a rigid side chain, distributed around a central five-membered ring. Unlike Ketoconazole (2), Itraconazole (3) and Saperconazole (4), which contain a central 1,3-dioxolane ring, SCH51048 contains a central tetrahydrofuran ring which confers increased efficacy.1e Construction of this 2,2,4trisubstituted ring with the required stereochemistry presented a synthetic challenge. The initial approach1c introduced chirality (88-92% ee) via a Sharpless-Katsuki epoxidation2 of the allyl alcohol 5. A series of further reactions yielded the desired R-triol 6, in which the enantiomeric excess had been enhanced to >98%. Tosylation of the two primary hydroxyls then gave ditosylate 7 (Scheme 1). The tetrahydrofuran ring could now be formed via nucleophilic stereoselective displacement of one of the two * Author to whom correspondence should be addressed. Phone: 908 298 7637; e-mail:
[email protected]. † Biotransformations Group. ‡ Chemical Process R & D. X Abstract published in Advance ACS Abstracts, September 15, 1997. (1) (a) 34th Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC), Orlando, Florida, 4-7th October 1994. (Abstract Nos. B16, F181, F183, F185, F187, F193, F195, F197 B179). (b) Blundell, P.; Ganguly, A. K.; Girijavallabhan, V. M. Synlett 1994, 263265. (c) Saksena, A. K.; Girijavallabhan, V. M.; Lovey, R. G.; Pike, R. E.; Desai, J. A.; Ganguly, A. K.; Hare, R. S.; Loebenberg, D.; Cacciapuoti, A.; Parmegiani, R. M. Biorg. Med. Chem. Lett. 1994, 4, 20232028. (d) Lovey, R. G.; Saksena, A. K.; Girijavallabhan, V. M. Tetrahedron Lett. 1994, 35, 6047-6050. (e) Saksena, A. K.; Girijavallabhan, V. M.; Lovey, R. G.; Desai, J. A.; Pike, R. E.; Jao, E.; Wang, H.; Ganguly, A. K.; Loebenberg, D.; Hare, R. S.; Cacciapuoti, A.; Parmegiani, R. M. Biorg. Med. Chem. Lett. 1995, 7, 127-132. (f) A preliminary report of some of this work has appeared previously: Saksena, A. K.; Girijavallabhan, V. M.; Lovey, R. G.; Pike, R. E.; Wang, H.; Ganguly, A. K.; Morgan, B.; Zaks, A.; Puar, M. S. Tetrahedron Lett. 1995, 36, 1787-1790. (g) Saksena, A. K.; Girijavallabhan, V. M.; Pike, R. E.; Wang, H.; Lovey, R. G.; Liu, Y.-T.; Ganguly, A. K.; Morgan, W. B.; Zaks, A. US Patent 5,403,937, April 4, 1995. (2) Katsuki, T.; Sharpless, K. B. J. Am. Chem. Soc. 1980, 102, 59745976.
S0022-3263(97)00992-4 CCC: $14.00
Chart 1
hydroxyls by the tertiary hydroxyl under basic conditions (path a or b in Scheme 1). However, under all tested conditions, it was the undesired 2R,4R-trans isomer 8b which predominated (60 trans:40 cis), and isolation of the desired 2R,4S-cis isomer 8a required careful chromatography (25-30% yield from the triol 6). Clearly the yield of the desired material could only be maximized by differentiation of the two primary hydroxyls prior to cyclization to the tetrahydrofuran ring. If the pro-S hydroxyl group of triol 6 could be blocked, tosylation and cyclization would then yield the required 2R,4S-cis isomer 8a.3 Since triol 6 is a 2-substituted-1,3-propanediol, enzymatic desymmetrization suggested itself as a convenient method to differentiate between the two primary alcohols.4 (3) Blocking the pro-S hydroxyl of triol 6, followed by tosylation, cyclization, deblocking, and tosylation would result in a five-step sequence to cis tosylate 8a. Alternatively, blocking the pro-R hydroxyl would require a seven-step sequence to 8a because of the extra steps to invert the 2R chiral center. Pro-R hydrolysis of the triol diester 10 or 11 would require six steps to 8a. (4) (a) Ramos Tombo, G. M.; Schar, H.-P.; Fernandez i Busquets, X.; Ghisalba, O. Tetrahedron Lett. 1986, 27, 5707-5710. (b) Boland, W.; Fro¨ssl, C.; Lorenz, M. Synthesis 1991, 1049-1072. (c) Danieli, B.; Lesma, G.; Passerella, D.; Riva, S. In Adv. Use Synthons Org. Chem. 1993, 1, 143-219. (d) Schoffers, E.; Golebiowski, A.; Johnson, C. R. Tetrahedron 1996, 52, 3769-3826.
© 1997 American Chemical Society
Enzymatic Desymmetrization of Prochiral 1,3-Propanediols
J. Org. Chem., Vol. 62, No. 22, 1997 7737
Scheme 1
Scheme 2
Results and Discussion A. Desymmetrization of Triol 6. The enzymatic acetylation of triol 6 in neat methyl acteate using porcine pancreatic lipase (PPL Sigma Type II) had been previously reported to display pro-R selectivity to give (R,R)-9 (Scheme 2), and six subsequent steps were required to arrive at the desired 2R,4S-cis tosylate 8a.1d,3 However, in our hands under the reported conditions, this acylation resulted in a mixture of diol, diacetate, and monoacetate with marginally useful diastereomeric excess (85-90% de). Lipases generally display the same prochiral selectivity for the acylation of 2-substituted-1,3-propanediols and for the hydrolysis of the corresponding diesters. As a result, products of opposite absolute stereochemistry may be obtained using the same enzyme under either acylating or hydrolytic conditions (Scheme 2).4b Consequently, the PPL-catalyzed hydrolysis of the diesters 10 and 11b was examined.3 However the reaction showed poor selectivity. Extensive overhydrolysis to triol 6 occurred and the monoester 9 was formed in poor de. Furthermore prochiral selectivity varied with reaction conditions.5 A screen of our enzyme collection was undertaken to identify enzymes showing high pro-S selectivity for the acylation of triol 6. Of the 86 commercially available enzyme preparations which were initially tested, 16 showed significant pro-R selectivity, using vinyl acetate (10 equiv) as acylating agent in EtOAc (Scheme 3). These gave the less desired monoacetate ((R,R-9), in some cases with moderate to high diastereomeric excess (Table 1). The 12 enzymes which showed the desired pro-S selectivity all suffered from low selectivity and/or rate of acylation. From these 12, Amano Lipase R (yielding the (R,S)-9 isomer in 78% de at 20% conversion) was chosen for a study of different solvents and different acetylating
agents, but no conditions were found to improve the diastereoselectivity. Since a highly selective pro-S enzyme had not been identified, further work was undertaken with Amano Lipase AK, the most selective of the pro-R enzymes. Two approaches were possible. (i) Acylation: Pro-R acetylation provided the monoacetate (R,R)-9 in high de,6 but a subsequent protection/ deacylation sequence was required to invert the newly formed R-chiral center, resulting in a seven-step sequence from R-triol 6 to the cyclized cis tosylate 8a.3 (ii) Hydrolysis: Since lipases generally display the same prochiral selectivity for acylation and hydrolysis reactions, Lipase AK catalyzed hydrolysis of the diacetate 10 would provide the desired monoacetate (R,S)-9 in an overall six-step sequence from R-triol to cis tosylate 8a. Initial hydrolyses in phosphate buffer yielded monoacetate (R,S)-9 with low de, so the effect of cosolvents, both miscible and immiscible, was examined. For miscible solvents the best de’s were observed with 10-20% THF; when increased to 50% THF no reaction was observed. With 5% MeCN or 10% acetone the reaction was slower and less selective. Of the ether solvents, the best results were obtained with 20% Et2O; 20% iPr2O and 30% TBME resulted in product of low de. The hydrolysis of a series of diesters was examined in 10-20%THF/50 mM KCl (pH 7.0) (Table 2). While the R,S-monoesters were formed with high de, the selectivity was poor and substantial amounts of overhydrolysis to triol occurred before complete consumption of starting material. Nevertheless, hydrolysis of the dibutyrate 11a was a convenient process, not only because of its high initial rate. Due to its low water solubility, the monobutyrate 12a could be partially purified after the enzymatic step; aqueous extraction of the crude product preferentially removed the more water soluble triol 6 which was formed by over hydrolysis. Furthermore, the chemical acylation and enzymatic hydrolysis reactions could be carried out in a one-pot reaction. When butrylation of the triol 6 in (5) The acylation of 6 and the hydrolysis of 10, 11b varied with enzyme lot and with the nature of the acyl group. Using Sigma PPL Type II, lot no. 23H0294 was more reactive than lot no. 41H0954 under acylation conditions. With both enzyme samples butyrylation (10 equiv of trifluoroethyl butyrate in TBME; 2× enzyme) was faster than acetylation (neat MeOAc; 3× enzyme); however the R,R-monoester was formed in poor de under both conditions. Under hydrolytic conditions (10 mM phosphate buffer; pH 7; 10% THF, 10% MeCN or 50% TBME as cosolvent) lot no. 41H0954 was more reactive than lot no. 23H0294. Hydrolysis of the diacetate 10 yielded the expected R,S-monoacetate 9 with generally poor de. For hydrolysis of the dibutyrate 11b under the same conditions, lot no. 23H0294 showed poor reactivity (10% conversion in 10%MeCN/buffer and 4% conversion in 50%TBME/buffer after 6h) but produced the expected R,S-monobutyrate 12b. In contrast, lot no. 41H0954 produced the R,R-monobutyrate under all hydrolytic conditions tested, albeit in poor yield. (6) After 2 h, a mixture of triol 6 (6.5 g), vinyl acetate (10 equivs), and lipase AK (6.6 g) in EtOAc (150 mL) showed a 6:9:10 ratio of 1:98: 1. After workup (R,R)-9 was obtained in 81% yield (97% de; [R]21D -17.7 (c 0.46, MeOH); lit.1d [R]21D -17.8 (c 1, MeOH))
7738 J. Org. Chem., Vol. 62, No. 22, 1997
Morgan et al. Scheme 3
Table 1. Acetylation of Triol 6. Screen Resultsa time, h
% triol 6
% monoacetate 9
% diacetate 10
% de
Pro-R Acylating Enzymes Amano Lipase AK (Pseudomonas sp.) Novo Lipozyme (Mucor miehei) Amano Lipase PS-30 (Pseudomonas fluorescens) Amano Lipase CES (Pseudomonas sp.) Sigma Porcine Pancreatic Lipase Type II Solvay Porcine Pancreatic Lipase Biocatalysts P. fluorescens
Enzyme
21 21 21 21 21 21 19
0 0 0 1 0 31 1
92 53 76 97 98 69 94
8 47 24 2 2 0 5
98.1 93.5 92.4 92.1 85.7 84.3 84.0
Pro-S Acylating Enzymes Amano Lipase R (Penicillium roquefortii) Interspex Bacterial Lipase/Esterase Amano Newlase A (Aspergillus. niger) Quest Acid Protease 200,000 Meito OF (Candida cylindracea) Sigma Protease Type IV (Streptomyces caespitosus) Amano Lipase G (Penicillium camembertii) Biocatalysts C. cylindracea Biocatalysts P. ciclopium Enzeco Fungal Acid Protease Sigma Protease Type XXIV Solvay AFP 2000
21 48 49 49 21 49 21 21 21 48 49 48
83 55 39 78 41 67 43 82 87 32 80 2
17 39 58 21 56 32 57 18 9 64 18 75
0 5 3 0 3 0 0 0 4 3 2 23
77.6 73.5 72.1 56.7 55.5 54.3 36.1 33.8 31.3 26.2 23.6 21.5
a
Conditions: triol 6 (0.11-0.16 mmol), vinyl acetate (7-43 equiv), EtOAc 1.0 mL, enzyme 8-180 mg, 250 rpm, rt. Table 2. Hydrolysis of Diesters 10 and 11a-d with Lipase AKa
10 11a 11b 11c a
R
cosolvent
time, h
% triol
% monoester
% diester
% de
initial rate µmol/h/mg enzyme
CH3 CH2CH2CH3 CH(CH3)2 (CH2)4CH3
10% THF 10% THF 20% THF 20% THF
48 3 49 6
17 5 10 16
57 57 81 64
26 38 9 20
97.0 96.7 95.9 >98
0.03 0.46 0.03 0.17
Conditions: diester (60-100 mg); enzyme (60-100 mg)(except for 10, 300 mg); 50 mM aqueous KCl; pH 7.0; rt.
THF was complete, the entire reaction mixture was diluted 10 fold with 50 mM KCl, the pH adjusted to 7.0, and the enzyme added. From this reaction mixture, the monobutyrate 12a was obtained (Scheme 4), essentially free of triol, in good yield (64%) with 99% de. (iii) Pro-S Acylation: As the hydrolysis of the dibutyrate 11b was being developed, an enzyme displaying useful levels of the desired pro-S selectivity was finally identified. Treatment of triol 6 with Novo SP435 in EtOAc or MeCN with vinyl acetate as acylating agent resulted in formation of (R,S)-9 with 94-97% de (Table 3). B. Improved Route to Tosylate 8a. Before either the acylation or hydrolysis routes to the key monoesters 9, 12 could be exploited, a more efficient route to cis tosylate 8a was identified.1f In this sequence (Scheme 5) the tetrahydrofuran ring is formed via an iodocyclization reaction from diol 13a. A 2S chiral center resulting
from a enzymatic desymmetrization of diol 13a could be used to control the stereochemistry of the iodocyclization reaction, forming predominantly the cis iodide 15 and leading ultimately to the desired (R,S)-tosylate 8a.7 From an initial screen of 53 hydrolases, 3 enzyme preparations were found to provide the monoacetate 14b with high enantiomeric excess (Table 4). The screen was subsequently extended, and a total of 205 commercial enzyme preparations have been examined under similar conditions; only one other enzyme preparation, Chirazyme L-2 (Boehringer-Mannheim),8 displayed high pro-S selectivity. (7) Acylation of the pro-S hydroxyl of 13a would result in a fivestep sequence to cis tosylate 8a (acylation, iodocyclization, triazole introduction, deacylation, and tosylation), while pro-R acylation would require two extra steps to invert the 2R chiral center. Pro-R hydrolysis of the diester 13a,b would result in a six-step sequence to 8a. (8) Both Novo SP435 and Boehringer-Mannheim Chirazyme L-2 are immobilized forms of Lipase Type B from Candida antarctica.
Enzymatic Desymmetrization of Prochiral 1,3-Propanediols Scheme 4
Scheme 5
Table 3. Acetylation of Triol 6 with Novo SP435a
Scheme 6
run solvent 1 2
J. Org. Chem., Vol. 62, No. 22, 1997 7739
EtOAc MeCN
time, % triol % monoacetate % diacetate min 6 9 10 % de 75 55
5 4
83 95
12 1
94.0 96.6
a Conditions: run 1: triol 64 mg, enzyme 15 mg, EtOAc1 mL, vinyl acetate 10 equiv. Run 2: triol 0.5 g, SP435 51 mg, MeCN 5 mL, vinyl acetate 2 equiv, rt.
Table 4. Acetylation of Diol 13a: Initial Screen Resultsa enzyme
time, min
% diol 13a
% monoacetate 14b
% diacetate 13b
% ee
pref
Lipase SP435 (Novo) Lipase CE (H. lanuginosa) (Amano) Lipase H. lanuginosa (Biocatalysts Ltd.)
60 95 220
0 0 1
83 97 98
17 3 1
97 99 98
pro-S pro-R pro-R
a
Conditions: diol 50 mg; enzyme 10-200 mg; vinyl acetate 10 equivs; toluene 1.0 mL; rt.
The availability of two enzymes with opposite prochiral selectivity allowed three approaches to the preparation of the desired S-monoester precursor. (i) Pro-R acetylation of diol 13a using Lipase CE gave the monoacetate (R)-14b with high chemical and optical yield. As the reaction profile indicated, kR1 >> kS1 or kS2 (Scheme 6). However, protection/deacylation was required to invert the chiral center, resulting in a sevenstep sequence from diol 13a to tosylate 8a.7 Since attempts to introduce a tetrahydropyranyl protecting group resulted in significant racemization, presumably via 1,3 acyl migration, this route was not pursued and was only used to prepare small quantites of (R)-14b. (ii) Lipase CE catalyzed pro-R hydrolysis of the diester would yield the S-monoester (kR-2 >> kS-2), for a sixstep sequence from diol 13a to tosylate 8a.7 While hydrolysis of the dibutyrate 13c produced the monobutyrate (S)-14c with high enantiomeric excess, the chemoselectivity was poor, and the reaction mixture consisted of a mixture of diol 13a, dibutyrate 13c, as well as the desired monoester (S)-14c. (iii) SP435 catalyzed pro-S acetylation of diol 13a provided (S)-14b directly for a five-step route to tosylate 8a.7 Because the prochiral selectivity of this enzyme is
lower than for Lipase CE (kS1/kR1 ≈ 20), optical purity was purchased at the cost of chemical yield, and a significant amount of the diacetate 13b had to be formed before the ee of the remaining monoacetate 14b reached useful levels (Scheme 6 and Figure 1).9 Nevertheless, because it provided the S-monoester directly under operationally simple conditions, the SP435 catalyzed acetylation of diol 13a was selected for optimization. Ironically, the first problem was not improvement, but reproduction of the favorable results from the initial screen. Diol 13a is prepared by reduction of the corresponding 2-substituted diethyl malonate and samples from different sources (LiAlH4 vs NaBH4/LiCl reduction, chromatographically pure vs crude product) displayed different reaction rates and selectivities. The different reaction rates were attributed to moisture in the samples, since the rate of diol consumption could be decreased by addition of small amounts of water, while molecular sieves increased the rate. More disturbing was the observation that, for certain samples, significant amounts of diol remained unconsumed, even though the reaction was run out to high diacetate formation.
7740 J. Org. Chem., Vol. 62, No. 22, 1997
Morgan et al. Scheme 7
Table 6. Product Distribution and Enantiomeric Excess at >94% Diol Conversion in Various Solventsa product distribution (%) diol/ vinyl enzyme monodiacetate ratio time diol acetate acetate % ee run solvent equiv (w/w) (min) 13a 14b 13b (S)
Figure 1. Acylation of 13a with Novo SP435: Diol, 50 mg; SP435, 10 mg; vinyl acetate, 10 equiv; toluene, 1.0 mL; rt. Table 5. Incomplete Enzymatic Acetylationa product distribution (%) vinyl diol/enzyme acetate ratio time diol monoacetate diacetate run solvent equiv (w/w) (min) 13a 14b 13b 1 2 3 4
toluene MeCN MeCN MeCN a
2.2 2.2 5.0 5.0
10 10 4.4 4.4
175 105 1080 1170
90 93 46 35
10 7 38 32
0 0 17 33
Conditions: diol, 0.5 g; SP435, 50 mg; solvent, 10 mL; 0 °C.
For one sample (Table 5, runs 1-3) the reaction was sluggish in both toluene and MeCN at 0 °C, with 95% with 0.97 with
